...The beauty of the adaptation of parts seems to me unparalleled. I should think or guess [that] waxy pollen was most differentiated. In Cypripedium, which seems least modified, and a much exterminated group, the grains are single. In all others, as far as I have seen, they are in packets of four; and these packets cohere into many wedge-formed masses in Orchis, into eight, four, and finally two. It seems curious that a flower should exist which could, at most, fertilise only two other flowers, seeing how abundant pollen generally is; this fact I look at as explaining the perfection of the contrivance by which the pollen, so important from its fewness, is carried from flower to flower. By the way, Cephalanthera has single pollen-grains, but this seems to be a case of degradation, for the rostellum is utterly aborted. Oddly, the columns of pollen are here kept in place by very early penetration of pollen-tubes into the edge of the stigma; nevertheless, it receives more pollen by insect agency. Epithecia [Dichaea] has done me one good little turn. I often speculated how the caudicle of Orchis had been formed. (600/2. The gradation here suggested is thoroughly worked out in the "Fertilisation of Orchids," Edition I., page 323, Edition II., page 257.) I had noticed slight clouds in the substance half way down; I have now dissected them out, and I find they are pollen-grains fairly embedded and useless. If you suppose the pollen-grains to abort in the lower half of the pollinia of Epipactis, but the parallel elastic threads to remain and cohere, you have the caudicle of Orchis, and can understand the few embedded and functionless pollen-grains. I must not look at any more exotic orchids: hearty thanks for your offer. But if you would make one single observation for me on Cypripedium, I should be glad. Asa Gray writes to me that the outside of the pollen-masses is sticky in this genus; I find that the whole mass consists of pollen-grains immersed in a sticky brownish thick fluid. You could tell by a mere lens and penknife. If it is, as I find it, pollen could not get on the stigma without insect aid. Cypripedium confounds me much. I conjecture that drops of nectar are secreted by the surface of the labellum beneath the anthers and in front of the stigma, and that the shield over the anthers and the form of labellum is to compel insects to insert their proboscis all round both organs. (600/3. This view was afterwards given up.) It would be troublesome for you to look at this, as it is always bothersome to catch the nectar secreting, and the cup of the labellum gets filled with water by gardener's watering. I have examined Listera ovata, cordata, and Neottia nidus avis: the pollen is uniform; I suspect you must have seen some observation founded on a mistake from the penetration and hardening of sticky fluid from the rostellum, which does penetrate the pollen a little.